A Deeper Look into Aleatoric and Epistemic Uncertainty Disentanglement
This work addresses the problem of trust in neural network predictions for practitioners and researchers, though it is incremental as it generalizes and evaluates existing methods rather than introducing a new paradigm.
The paper investigates the disentanglement of aleatoric and epistemic uncertainties in neural networks, revealing unexpected interactions between them and finding that ensembles provide the best disentangling quality, with recommendations such as using over 100 samples for reliable results.
Neural networks are ubiquitous in many tasks, but trusting their predictions is an open issue. Uncertainty quantification is required for many applications, and disentangled aleatoric and epistemic uncertainties are best. In this paper, we generalize methods to produce disentangled uncertainties to work with different uncertainty quantification methods, and evaluate their capability to produce disentangled uncertainties. Our results show that: there is an interaction between learning aleatoric and epistemic uncertainty, which is unexpected and violates assumptions on aleatoric uncertainty, some methods like Flipout produce zero epistemic uncertainty, aleatoric uncertainty is unreliable in the out-of-distribution setting, and Ensembles provide overall the best disentangling quality. We also explore the error produced by the number of samples hyper-parameter in the sampling softmax function, recommending N > 100 samples. We expect that our formulation and results help practitioners and researchers choose uncertainty methods and expand the use of disentangled uncertainties, as well as motivate additional research into this topic.